statgraphics 18®

Whats’ new in Version 18?

DATA MANAGEMENT 1. Big Data – new file format for managing data with many millions of rows. 2. Censored numeric data type – for representing left, right and interval-censored data. 3. Create data and code columns – for consolidating multiple data columns. 4. Currency data type – for representing currency in dollars, Euros, pounds or yen. 5. New operators – 17 new operators for use in Statgraphics expressions. 6. Replace censored values – specification of replacement values for censored data.

GRAPHICS 1. Contour plots – new option to label each contour line. 2. Demographic maps – now supports SHP boundary files. 3. Diamond plot – dot plot with diamond displaying confidence interval for the mean. 4. Donut chart – alternative to piechart. 5. Heat map – shows the distribution of a quantitative factor by 2 categorical factors. 6. Hexagon plot – automatic replacement of scatterplots with hexagons for Big Data. 7. Likert plot – displays survey data recorded on a Likert scale. 8. Regression models – shading of area between confidence and prediction limits. 9. Ribbon plot – new type of plot for displaying response surfaces. 10. Time series baseline plot – for plotting data with upper and lower warning limits. 11. Tornado and butterfly plots – for comparing 2 samples of attribute data.

SYSTEM OPERATION 1. Installation – new option to deactivate program and move to new computer. 2. Network installation – new option to check out individual seats. 3. R interface – improved configuration of interface to the R language. 4. Repeat analysis by – replication of statistical procedures for multiple data columns.

DESIGN OF EXPERIMENTS & STATISTICAL PROCESS CONTROL 1. Attribute capability analysis Statlet – classical and Bayesian estimation. 2. Capability control charts – for Cp, Cpk, rate & proportion of nonconforming items. 3. Capability control chart design – determination of adequate sample size. 4. Definitive screening designs – new designs for fitting models with quadratic terms. 5. Multivariate capability analysis – new bootstrap confidence intervals. 6. Multivariate tolerance limits – using elliptical regions or Bonferroni approach. 7. X-bar and R chart – now available for subgroups with more than 100 observations.

NEW STATLETS 1. Demographic map visualizer – dynamic visualization of changes by location. 2. Population pyramid – dynamic visualization of comparative population changes. 3. Sunflower plot – displays Big Data with sunflowers replacing overlapping points. 4. Trivariate density – for displaying multivariate distribution of 3 variables. 5. Violin plot – combines box-and-whisker plot with nonparametric density estimate. 6. Wind rose – dynamic visualizer of changes in wind speed and direction over time. statgraphics 18®

OTHER NEW STATISTICAL PROCEDURES 1. Classification and regression trees – new machine learning algorithms. 2. Distribution fitting (arbitrarily censored data) – left, right and interval censoring. 3. Equivalence & noninferiority tests – for comparing two independent samples, two paired samples, one mean against a target, or a 2x2 crossover study. 4. Multidimensional scaling – displays multivariate data in a low dimensional space. 5. Multivariate normal random numbers – generates data with specified correlations. 6. Multivariate normality test – tests if data come from a multivariate normal distribution. 7. Orthogonal regression – fits models with errors in both Y and X. 8. Text mining – document scanning and construction of word clouds. 9. X-13ARIMA-SEATS Seasonal Adjustment – for time series data.

CHANGES TO EXISTING PROCEDURES 1. Automatic forecasting – Ljung-Box test for testing residual autocorrelations. 2. Bivariate density plot – overlays points on bivariate density contours. 3. Box-and-whisker plots – optional diamond to display confidence interval for mean. 4. Capability analysis for variables – new option to fit Johnson curves. 5. Chi-square tests – specification of minimum frequency before combining classes. 6. Monte Carlo general simulation models – new Sensitivity Tornado Plot. 7. Monte Carlo simulation of ARIMA models – option to specify starting values. 8. Subset analysis – new percentile plot and table of percentiles by group.